Tyler Duffey as Object Lesson
It’s also, to be clear, still an article about baseball! More specifically, it’s about Tyler Duffey. He’s “breaking out” this year, in that he’s faced 16 batters and struck out 10 of them. That sample size? It’s too small to really say anything. Take a look at our handy sample size tool, and you’ll realize it in no time. And yet, we write these articles. Maybe it’s this piece on Chaz Roe, or this one on Tommy Kahnle getting good, or this one on Nick Anderson striking everybody out — over the years, they’ve become FanGraphs staples. How?
Tyler Duffy is a great example of this. By pretty much any conceivable measure, he’s the best reliever in baseball so far this year. FIP? Tied for first with three guys who have only thrown an inning. xFIP? Second behind Colin Rea, one of the aforementioned one-inning wonders. Strikeout rate? First? Walk rate? Well, he hasn’t walked anybody, so that’s a yes.
That’s all meaningless, because again, he’s faced 16 batters. But you know what isn’t? The 238 batters he faced in 2019, and he was great then too. He had a 34.4% strikeout rate, 17th among relievers with 50 innings pitched. He had a dazzling 2.94 xFIP, good for 14th. And despite some iffy home run luck, he posted 1.2 WAR, phenomenal for a reliever who didn’t get the benefit of entering games in high-leverage spots (his 0.92 inLI was lower than every reliever who out-WAR’ed him other than John Brebbia).
So there’s the setup. I’m using this small current sample plus a hidden larger sample to tell you, confidently, that this guy you might not have heard of is really good. Why should you believe me, aside from the fact that I’m writing on this fancy blog? It’s time to throw in some evidence. You’re a discerning sort, anonymous FanGraphs reader. You’re definitely going to want some evidence.
Good news again! Baseball is lousy with evidence these days. You can’t look at five innings of play without drowning in data. Spin rates, pitch movement, pitch location, batter handedness, release points — it’s all out there. Want to look at a scatterplot of the pitches Duffey throws? Here! It’s not illuminating, but it’s easy to make:
How about something more useful, though? Duffy was bad in 2018, so-so in 2017, and a decidedly medium starter in 2016. Why did he break out in 2019? What the heck, I’ve got you covered there too. Here’s another trope: he changed his pitch mix and found success. He even talked about it with David Laurila last year. He started throwing his curve faster and with less East/West break, trusted it more, and fortune followed:
| Year | HMov (in) | VMov (in) | Speed (mph) | SwSt% | Usage Rate |
|---|---|---|---|---|---|
| 2018 | 5.5 | -8.1 | 80.7 | 16.3% | 33.8% |
| 2019 | 3.4 | -7.2 | 82.6 | 16.7% | 45.9% |
| 2020 | 1.0 | -5.5 | 82.8 | 30.2% | 58.9% |
Don’t pay too much attention to the specifics, because they’re not incredibly important — particularly the gobsmacking 2020 swinging strike rate. I’m dazzling you with small samples, remember? He’s only thrown 43 breaking balls this year. The usage rate is important, though: one of the big lessons I’ve taken from modern pitching breakthroughs is that throwing your best pitches more often is a great plan, and Duffey excised a sinker and changeup he’d used a combined 30% of the time between 2018 and 2019, replacing them with the curve and his four-seam fastball.
At this point in the article, your attention is probably flagging. That’s partially because of the contrived structure, but it’s also because it’s hard to read a thousand words about a guy who has thrown 70 pitches this year and keep your focus. Now it’s time for another article staple: the swinging strike GIF. Goodness, it’s glorious:
And hey, that curveball sets up a nice fastball, too. Batters have to think about the breaking ball — he’s throwing it more than half the time, after all — and that leaves them vulnerable to the cheese:
Those pitches are objectively fun to watch. They’re objectively good, too; a major league hitter swung and missed at them. That doesn’t mean that all his pitches are good, or that his curve consistently makes batters look that bad. But in this case, trust me, it’s good! He was already in the top 20% of baseball for curveball whiff rate last year, and he even has another mode to it. He can, as he alluded in his conversation with Laurila, take something off of it and loop one in there to steal a strike from batters geared up for the faster model:
Now that you’ve watched Duffey excel, it’s time for the next article key: I redirect you to the main point. Duffey’s great now. He was already great, and now you’re learning about it. He has that sweet curveball, he’s leaning into it, and that’s that. The Twins already know. His most recent appearance was to protect a one-run lead in the eighth, and given that he’s not taking the ninth inning job from Taylor Rogers anytime soon, that’s as good as it gets in Minnesota.
By this point, it’s time to throw in a truism about either baseball in general or Duffey in particular. How about this one? In modern baseball, you can’t have too many relievers. The Twins were one of the best teams in baseball last year, and they likely are again this year. In the rarefied air of the playoffs, you don’t just need one guy like Taylor Rogers. You need a deep bench — literally in this case — of guys who can pitch the sixth, seventh, and eighth innings.
The Twins had that last year, though they didn’t show it in the playoffs — vagaries of a three-game series and all. And Duffey was a key part of it, particularly by the end of the year. Last September, he threw 11.2 innings. He struck out 51.2% of the batters he faced and walked none. His second best month? August, with a 40.5% strikeout rate, though he paired it with an unsightly 11.9% walk rate. He gave up a single homer and two earned runs over those two months combined.
So yeah, I’m writing an article about a guy who has thrown five innings this year. They were five pretty great innings, though, and his last 22.1 of 2019 weren’t too shabby either. He was bad in the playoffs, but what Twin wasn’t? If we can’t forgive him that, the whole team is irredeemable. That slight blip aside, Duffey has been one of the very best relievers in baseball for the last three months worth of baseball. It’s still a tiny sample, but it’s hardly five innings.
With that grand flourish, it’s time for the conclusion of this article about both an exciting reliever and the art of writing about exciting relievers. Duffey is great, and articles about tiny sample sizes are a cinch. All you have to do is cheat! Take some old stats and use them to bulk up your data. Anyone could copy down the list of strikeout rates this year and point Duffey’s name out. Stealing stats from 2019 to make an argument about 2020, though? Picasso would approve.